1. Home
  2. ⏣ General Metalworking
  3. The problem with precision

The problem with precision

My Fadal mill showed me by the cuts it was making that the spindle needed to be trammed. I have a lot of good machine shop inspection equipment. I can easily, repeatably, and accurately measure down to 50 millioniths of an inch. Not so easily but accurately and reapatably down to 20 millionths. I have several standards for lengths. My best set are the ceramic gauge blocks. The maximum deviation from nominal size of any of the blocks is +3 millionths of an inch except the 4 inch block which is +4 millionths. Most are only + 2 or less millionths. I have indirectly seen temperature changes by measuring lengths. Anyway, back to the Fadal. After leveling the mill according to the Fadal manual instructions with a precision level to better than .0005 per foot I checked the table flatness in the six spots corresponding to the spindle retaining bolt locations, again following the Fadal instructions. Then, using a tool that Fadal recommends, I set about to tramming the spindle. The Fadal manual says to tram to better than .001 in 12 inches and that Fadal Techs usually tram the spindles to .0005. So I figured I would make mine better. Finally, after fussing with the damn thing for hours and hours I got the spindle tram to just slightly better than .0004 in 12. The problem is that the error is so magnified by the measuring tools that .0004 looks like a huge amount. And barely touching anything moves the indicator. I checked and re-checked and so on for the next day just to make sure the machine wasn't going to settle and change the tram. Since I can see tiny changes it makes me want to get the machine even closer. But it's not practical. I am sure that after making some cuts the spindle tram will change. And I have the tools and knowhow to measure the tram error. So I have to resist the temptation to check the tram because I know it will never end. It's the same with parts I make. I check 'em, see an error, change an offset, see another error, change another offset, and all the while the parts are within .001 and the tolerance is + or - .005. It's like a disease. It also affects me when I start to optimize programs to shave of smaller and smaller increments of time. At least now that the spindle is in proper tram the cuts look much better. I am milling some brass door furniture that has large flat surfaces inside pockets. These surfaces will be polished to a high shine so the cutter marks must be able to be removed easily. So it's a purely cosmetic thing but since the pockets can't be sanded on a surface plate to remove cutter marks the cutter marks must be removable with only tripoli or E5 emery on a buffing wheel or point. After tramming the head the parts are now coming out the way I want. Eric

Reply to
etpm
Loading thread data ...

Hmm. You're looking for Mikron or Kugler accuracy from a Fadal. I'm surprised you're getting it that close. Don't sneeze!

Reply to
Ed Huntress

The model I have has the linear ways. Though not as heavy duty as box ways they don't wear the way box ways do. Since the mill is back in pruduction it has probably changed already and I am going to resist checking it again unless I see another problem. Unbelievably the 4th axis that came with the machine is out of square. This is the stock Fadal 4th axis. I bought the machine used so I expected some error somewhere but the amount of out of squareness in the 4th axis is huge for this kind of machine. When used with the rotary table surface parallel to the Z axis I have to put .006" shims under one edge of the thing in order for the table to be square to the mill table. The back surface is parallel to the table surface. I thought that maybe the table spindle is adjustable in the casting that holds it but there is no adjustment. The holes machined in the casting for the spindle are off. Too bad the rotary axis spindle can't be trammed like the main machine spindle can. One of these days I am going to set the 4th axis on the surface grinder and square it up after I rough it in on the mill. Eric

Reply to
etpm

Repeat after me:

Measure with digital calipers.

Mark with chalk.

And cut with an axe.

All better now?

Reply to
Larry Jaques

Well, don't tell them I told you this, but Fadals typically are pretty rough-and-ready machines, and are not known for accuracy. A guy I knew who used three of them to machine EDM electrodes, when asked why he used Fadals, told me "because the graphite dust is going to wear out any mill, and I can scrap my Fadals every two years without feeling bad about it." I sold him a Roku-Roku with and enclosed air-blast plus vacuum, and it solved that problem. It also gave him about five times the accuracy.

So, what are your ceramic gage blocks, Mitutoyo?

Reply to
Ed Huntress

I know that Fadals are not the most accurate or fastest or beefiest. My 16 x 30 inch travel mill weighs about 5500 or 6000 lbs. and has a

10 HP. spindle while my Miyano lathe with 13.5 inches of Z axis travel and 5.75 of X axis travel has a 15 HP. spindle and weighs 8000 lbs. But I still make good parts with it, to tight tolerances, and it has 4 axes. It also has the best control of any machine I have ever used. It is the most intuitive. And tool setting is fast and easy. And the gauge blocks are Mitutoyo. The ceramic ones. My shop set is an 81 piece that I got from SPI. I think. I've had them for something like 30 years. Eric
Reply to
etpm

Yeah, they're great. I have a miscellaneous assortment of them, not a set, that I a...acquired when I was doing photography for them. They'd give me a digital mike or a surface plate to photograph, and then they'd tell me not to bring it back, because they'd already written it off as advertising expense.

I wrote a whole string of publicity articles and other material for them about the Cera Blocks when they came out. They were my second-biggest client, after Makino.

Reply to
Ed Huntress

The system I use is I put a wide chunk of thick aluminum sheet on the table, and set up a CNC program to mill a circular groove it it, with an end mill. So, the table orbits around, making an over 6" diameter circle. Then, it moves to the center of that circle. I can mount a dial test indicator in the spindle and sweep it around the groove. This allows me to detect tram errors relative to the ACTUAL PLANE of X-Y motion, not the top of the table, which might not be the same.

Why do you think the tram will change. Once I got mine set up, it hasn't changed measurably in years. My table does not actually have a "plane" of motion, as all the ways are a bit worn, and that circular ring is actually a sort of saddle shaped.

Jon

Reply to
Jon Elson

To tram the head the first thing to do is check the flatness of the table in six spots corresponding to the spindle bolts. Then you take into account the flatness measurements when you tram the head. I was being a little facetious when I said the spindle would probably move. I spent a long time with the torque wrench and shims getting the spindle just as close as I could. On the other hand, when I bought the machine the spindle tram was good. After setting it up in my shop I leveled the mill and the tram was still good. But after a couple years or so the level changed and so did the tram. Eric

Reply to
etpm

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.